repo-setup

Pass

Audited by Gen Agent Trust Hub on Aug 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes multiple shell scripts located in the plugin's bootstrap directory (check-env.sh, init-profile.sh, sync-labels.sh, scaffold-verify.sh) to perform environment diagnostics and initialize repository metadata.\n- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests data from untrusted repository files to populate configuration templates.\n
  • Ingestion points: Reads package.json, turbo.json, biome.json, and tsconfig.json to infer toolchain commands and project structure.\n
  • Boundary markers: None identified; the skill does not explicitly instruct the agent to ignore instructions embedded within the ingested configuration files.\n
  • Capability inventory: The skill can execute shell scripts, write to the filesystem, and interact with the GitHub CLI.\n
  • Sanitization: There is no evidence of sanitization for the data extracted from configuration files before it is interpolated into the repo-profile.md skeleton.\n- [EXTERNAL_DOWNLOADS]: The skill preflight script identifies missing CLI dependencies and suggests downloading them from established external sources.\n- [REMOTE_CODE_EXECUTION]: The skill recommends using the official bun curl-to-bash installer to resolve environment issues identified during the preflight stage.\n- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill interacts with GitHub APIs via the gh CLI to synchronize repository labels and verify authentication status.\n- [DYNAMIC_EXECUTION]: The skill generates and subsequently runs TypeScript verification scripts using the bun runtime.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 15, 2026, 10:26 PM
Security Audit — agent-trust-hub — repo-setup